Hi Guys, In the past timegpt was always pretty spot on but in my latest forecast I am struggling to understand a forecast. Light Blue is are the actual values, darker blue is the timegpt forecast. Everything makes sense except the forecast for 12-2024 (forecasted 106). As you can see from the previous years there was always a pretty big spike. December was almost every time 80% over November but the forecasted is way lower this time and I can't understand why timegpt isn't able to figure this out. This is my forecasting code, no other modifications to my data:

MonthlyForecastDF = timegpt.forecast(df=MonthlyDataDF, h=9, freq='MS', time_col='Date', target_col='MaxSales')

Hello, You could try fine-tuning the model using

fintetune_steps=10

for example. Also, you can specify the loss used for fine-tuning using

finetune_loss

. You can choose from mae, mse, rmse, mape and smape. I hope this helps!